Just further proof that context is the real moat, not intelligence. All the models are already converging to be equally intelligent and that will only continue.

Just further proof that context is the real moat, not intelligence. All the models are already converging to be equally intelligent and that will only continue. GPT 5.4 / Opus 4.6 are the first two models I’ve used where I’m like, yeah, with the right spec/context they can pretty much do anything. The “bundle” or “context” is the value.